Objective: The purpose of this study was to investigate whether motor imagery training (MIT) combined with physical training could have a positive influence on brain waves in stroke patients. Design: Randomized controlled trial. Methods: Twenty-four survivors of a first-time stroke participated in the study. All patients were randomly assigned to one of two groups: (1) MIT group (n=12) and (2) control group (n=12). The MIT group participated in a total of 30 minutes of motor imagery (MI) on each of 18 tasks related to their daily living over the course of four weeks. Provided video images before training of MI were discussed and decided by a professor of occupational therapy, occupational therapists and physical therapist. For the MIT group, during each 30 minute interval, MI was conducted for 20 minutes and participants underwent physical training for 10 minutes. Both groups received conventional therapies for 30 minutes, five times per week for four weeks. After the four-week intervention period, we compared the outcomes of the brain wave recordings between the two groups. Results: When compared with the control group, relative alpha waves for frontal lobe F3 and F4, and relative beta waves for frontopolar Fp2, F3, and F4 were more significantly improved in the experimental group (p<0.05). Conclusions: MIT combined with physical training has a positive influence on brain waves in stroke patients. These results suggest that the MIT is feasible and beneficial for improving brain activity for persons who have had a stroke.